US9591320B2 - Context and bypass encoding video - Google Patents
Context and bypass encoding video Download PDFInfo
- Publication number
- US9591320B2 US9591320B2 US13/184,226 US201113184226A US9591320B2 US 9591320 B2 US9591320 B2 US 9591320B2 US 201113184226 A US201113184226 A US 201113184226A US 9591320 B2 US9591320 B2 US 9591320B2
- Authority
- US
- United States
- Prior art keywords
- encoded
- context
- bin
- array
- transform coefficient
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active, expires
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/436—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation using parallelised computational arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/13—Adaptive entropy coding, e.g. adaptive variable length coding [AVLC] or context adaptive binary arithmetic coding [CABAC]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
- H04N19/139—Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/46—Embedding additional information in the video signal during the compression process
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/91—Entropy coding, e.g. variable length coding [VLC] or arithmetic coding
Definitions
- Embodiments of the present invention generally relate to a method and apparatus for parallel context processing techniques for high coding efficiency entropy coding, which may be used in the video coding standard High Efficiency Video Coding (HEVC).
- HEVC High Efficiency Video Coding
- Embodiments of the present invention relate to a method of encoding video information comprising receiving a series of syntax elements comprising an array of transform coefficients related to a block of an image, determining a context for at least a first bin and a second bin of each significant transform coefficient in the array, context encoding the at least a first bin and a second bin of each significant transform coefficient in the array, bypass encoding a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array, bypass encoding a sign indicator for each significant transform coefficient in the array, and constructing a bit stream.
- FIG. 2 is an embodiment of a flow diagram depicting PIPE/V2V coding
- FIG. 3 is an embodiment of a syntax element partitioning
- FIG. 4A is an embodiment of a flow diagram depicting a parallelization of context processing for significance map utilizing speculative computing at each bin;
- FIG. 5 is an embodiment of a flow diagram depicting a method for context processing tree for level coding in AVC
- FIG. 6 is an embodiment of a flow diagram depicting context processing tree for levels when SIGN is coded in separate bin-plane.
- FIG. 1 is an embodiment of a CABAC block diagram.
- the serial nature in CABAC comes from the following three blocks, a binarizer, a context modeler and a binary arithmetic coder.
- binarizer bins from many syntax elements, such as, coefficient levels and motion vector differences are coded using variable length coding such as unary coding and exp-Golomb coding.
- Variable length codes are inherently serial in nature.
- the context modeler the serial dependency comes about since the probability used in the context model for coding the next bin is updated depending on the current bin value. If the current bin value is Least Probable Symbol (LPS), the probability is increased and if the current bin value is Most Probable Symbol (MPS), the probability is decreased.
- LPS Least Probable Symbol
- MPS Most Probable Symbol
- context index selection process Another source of serial dependency is the context index selection process, where the context index of bin may be determined by the value of previously coded bins.
- the arithmetic coding uses interval subdivision. The range, value, offset used to determine the interval on [0, 1] that uniquely identifies the coded stream of bin values are updated in a serial fashion as and when bins get encoded/decoded.
- the parallelism proposed may be broadly classified into three categories: (1) Bin-level parallelism, which parallelizes the BAC, (2) Syntax element-level parallelism, which parallelizes the BAC, the context modeler, and the binarizer and (3) Slice-level parallelism.
- NBAC N-bins/cycle coding
- the contexts for N-bins are calculated through the use of conditional probabilities.
- the binarizer and context modeler were basically the same as in CABAC of AVC.
- coding schemes are determined variable-to-variable length for coding of the bins.
- PIPE/V2V coding scheme the bins are coded using a parallel bin encoding scheme as shown in FIG. 2 .
- FIG. 2 is an embodiment of a flow diagram depicting PIPE/V2V coding.
- the NBAC, PIPE, V2V schemes reduces serial dependency in the BAC block.
- the serial dependency in the context modeler and binarizer still remain. So, the effective throughput increase that can be achieved in entropy coding is limited.
- techniques for parallelization of context processing may be utilized.
- FIG. 4B is an embodiment of a flow diagram depicting a parallelization of context processing for significance map utilizing speculative computing at a fifth bin. If all of the significant coefficient flag is zero, then the last significant coefficient flag is not transmitted.
- Table 1 shows the distribution of bins used by different syntax element types as a percent of total bins for a LCTB.
- the bin distribution was obtained by measuring bins in bitstreams generated, for example, by TMuC-0.1 using cfg files in cfp-fast directory. Shown in Table 1 is the distribution of bins used by different syntax element type as a percent of total bins for a LCU.
- FIG. 5 is an embodiment of a flow diagram depicting a method for context processing tree for level coding in AVC.
- the context processing tree that needs to be pre-calculated at each bin to achieve 4 ⁇ parallelism in context processing of level in AVC.
- the context processing that happens at every SIGN node is wasteful since SIGN is coded in bypass mode.
- Table 3 shows the distribution of level of coefficients obtained by measuring levels in bitstreams generated by, for example, a TMuC-0.1 using cfg files in cfp-fast directory.
- FIG. 6 is an embodiment of a flow diagram depicting context processing tree for levels when SIGN is coded in separate bin-plane.
- the context processing tree for levels when sign is coded in separate bin-plane.
- context processing efficiency is 100%. This is also illustrated in FIG. 7 where all sign bins (i.e. Coeff Sign Bins) are coded on separate bin plans; this is referred to as Coeff Sign PCP.
- the first two bins in the coefficient level are context coded.
- the rest of the bins such as, coefficient sign bins and Golomb-Rice+Exp-Golomb (GR-EG) binarized bins, are bypass coded.
- the second bin in the absolute value of the coefficient minus 1 i.e. Coeff Level BinIdx 1
- the Coeff Sign Level can be interleaved or be on a separate bin-plane with GR-EG bins.
- FIG. 7 is an embodiment of a proposed approach on order of syntax elements.
- FIG. 7 illustrates a data ordering based on Coeff Level BinIdx 0 PCP, Coefficient Sign PCP, and Coeff Level BinIdx 1 PCP.
- Bypass coded bins are Coefficient Sign & GR-EG bins.
- the first row shows original ordering used in H.264/AVC.
- c 0 and sign can be placed in partitions that can be coded in parallel with the other bins.
- the new coefficient level binarization and coding introduced in HM-3.0 is shown in the third row.
- Coeff Level BinIdx 1 PCP is shown in the fourth row.
- c 0 , c 1 and sign+GR-EG bins can be placed in partitions that can be coded in parallel with the other bins.
- sign and GR-EG bins i.e. exp-golomb and golomb rice bins of coeff
- bypass bins can be coded faster than context coded bins.
- many bypass bins can be coded in a cycle which can increase the throughput of the CABAC.
- Coeff Level BinIdx 1 PCP all bypass coded bins for coefficients in a given TU are grouped together which increases throughput impact of parallel bypass bins processing.
- Variants of this approach include separating GR-EG +sign bins from the Coeff Level BinIdx 0 and Coeff Level BinIdx 1 , but keeping the GR-EG +sign bins interleaved and keeping the Coeff Level BinIdx 0 and Coeff Level Bin Idx 1 bins interleaved as shown in proposal #2 in FIG. 11 .
- This eliminates the additional of loops required to separate the Coeff Level BinIdx 0 from Coeff Level BinIdx 1 , and Coefficient Sign from RG-EG bins.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computing Systems (AREA)
- Theoretical Computer Science (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
A method and apparatus for encoding and decoding video data. Including context encoding or decoding at least two bins of each significant transform coefficient in an array of transform coefficients and bypass encoding or decoding a sign indicator for each significant transform coefficient in an array related to a block of an image.
Description
This application claims benefit of U.S. provisional application Ser. No. 61/499,852, filed Jun. 22, 2011, U.S. provisional patent application No. 61/499,852, filed Jun. 22, 2011, and U.S. provisional patent application Ser. No. 61/364,593, filed Jul. 15, 2010, are hereby incorporated by reference.
Field of the Invention
Embodiments of the present invention generally relate to a method and apparatus for parallel context processing techniques for high coding efficiency entropy coding, which may be used in the video coding standard High Efficiency Video Coding (HEVC).
Description of the Related Art
Context-Adaptive Binary Arithmetic Coding (CABAC) is one of two entropy engines used by the existing video coding standard AVC. CABAC is a method of entropy coding that provides high coding efficiency. Processing in CABAC engine is highly serial in nature. Consequently, in order to decode high bit rate video bit-streams in real-time, the CABAC engine needs to be run at extremely high frequencies which consumes a significant amount of power and in the worst case may not be feasible.
Therefore, there is a need for an improved method and/or apparatus for parallel context processing techniques for high coding efficiency entropy coding in HEVC.
Embodiments of the present invention relate to a method and apparatus for parallel context processing for example for high coding efficient entropy coding, such as, HEVC. The method comprising retrieving syntax element relating to a block of an image, grouping at least two bins belonging to similar context based on the syntax element, and coding the grouped bins in parallel.
Embodiments of the present invention relate to a method of encoding video information comprising receiving a series of syntax elements comprising an array of transform coefficients related to a block of an image, determining a context for at least a first bin and a second bin of each significant transform coefficient in the array, context encoding the at least a first bin and a second bin of each significant transform coefficient in the array, bypass encoding a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array, bypass encoding a sign indicator for each significant transform coefficient in the array, and constructing a bit stream.
Embodiments of the present invention relate to an encoder comprising a binarizer operable to binarize syntax elements, a context modeler operable to determine a context for a portion of the binarized syntax elements, including at least a first bin and a second bin of each significant transform coefficient in an array of transform coefficients related to a block of an image, and a binary arithmetic coder. The binary arithmetic code comprising a context coding engine operable to context encode the at least a first bin and a second bin of each significant transform coefficient in the array and a bypass coding engine. The bypass coding engine operable to bypass encode a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array and bypass encode a sign indicator for each significant transform coefficient in the array.
So that the manner in which the above recited features of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to embodiments, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.
In some embodiments of parallel entropy coding tools, the parallelism proposed may be broadly classified into three categories: (1) Bin-level parallelism, which parallelizes the BAC, (2) Syntax element-level parallelism, which parallelizes the BAC, the context modeler, and the binarizer and (3) Slice-level parallelism.
A N-bins/cycle coding (NBAC) encodes and decodes N-bins/cycle to achieve N-fold improvement in throughput. The contexts for N-bins are calculated through the use of conditional probabilities. In some HEVC embodiment, the binarizer and context modeler were basically the same as in CABAC of AVC. However, coding schemes are determined variable-to-variable length for coding of the bins. There are two flavors of the scheme: (1) PIPE and (2) V2V. The main difference between the two is the context probabilities are quantized to 12 levels in PIPE and to 64 in V2V. In PIPE/V2V coding scheme, the bins are coded using a parallel bin encoding scheme as shown in FIG. 2 . FIG. 2 is an embodiment of a flow diagram depicting PIPE/V2V coding.
Some embodiments that utilize schemes that interleaves the V2V code words from different partial bitstreams into a single bitstream. As a result, a throughput increase of 6× for PIPE in hardware is possible. Such embodiments usually cause an estimated throughput increase of 3× in BAC stage for PIPE hardware implementation for both the parallel and serial versions of PIPE. Since PIPE uses 12 bitstream buffers and V2V uses 64 bitstream buffers, PIPE is usually utilized more often than V2V from a complexity purpose. However in both cases, there is no estimated overall throughput improvement in the entropy coder due to serial bottlenecks in context processing and binarization.
The NBAC, PIPE, V2V schemes reduces serial dependency in the BAC block. However, the serial dependency in the context modeler and binarizer still remain. So, the effective throughput increase that can be achieved in entropy coding is limited. Hence, techniques for parallelization of context processing (PCP) may be utilized.
In syntax element partitioning, syntax elements such as macroblock type, motion vectors, transform coefficients, significant coefficient map etc. are divided into N groups and each group is coded separately. The context selection and adaptation within a group happens in parallel leading to a potential N-fold speed up in context modeler if the various partitions are balanced in terms of the number of bins they process. In practice, the various partitions are not balanced and the throughput improvement is less than a factor of N.
Syntax element partitioning results showed throughput improvement and BD-Rate. In this embodiment, significance map coding is carried out in AVC CABAC. In such an embodiment, the last significant coefficient flag is transmitted when the related coefficient is determined to be significant. The coefficient is the output of a block after transform and quantization. Also, a coefficient is significant when it has value that is non-zero.
This technique introduces serial dependency in decoding of significance map. When throughput improvement is needed, speculative computation are performed at every bin. Such computations leads to complex logic, as shown in FIG. 4A . FIG. 4A is an embodiment of a flow diagram depicting a parallelization of context processing for significance map utilizing speculative computing at each bin. Speculative computation at every bin also results in increased power consumption.
Significance map coding are parallelized by transmitting the last significant coefficient flag once per certain number of bins. For example, FIG. 4B is an embodiment of a flow diagram depicting a parallelization of context processing for significance map utilizing speculative computing at a fifth bin. If all of the significant coefficient flag is zero, then the last significant coefficient flag is not transmitted.
Such an embodiment reduces the number of last bins that need to be transmitted, but it increases the number of significant bins that need to be transmitted. However, there is about a 5% overall reduction in the number of significance map bins that need to be processed. Our algorithm parallelizes about 21.65% of the bins for largest coding unit (LCTB).
Table 1 shows the distribution of bins used by different syntax element types as a percent of total bins for a LCTB. The bin distribution was obtained by measuring bins in bitstreams generated, for example, by TMuC-0.1 using cfg files in cfp-fast directory. Shown in Table 1 is the distribution of bins used by different syntax element type as a percent of total bins for a LCU.
TABLE 1 | |||
Bins used | Average number | ||
per syntax | of bins | ||
SigMap | 21.65% | ||
SigLast | 8.35% | ||
LevelAbs | 16.67% | ||
LevelSign | 9.92% | ||
The coefficient coding is usually carried out in AVC CABAC. The context used for the absolute value of the coefficient minus one , known as the coefficient level (1) depends on the position of the bin. Thus, when the binIdx is 0(i.e. first bin of the coefficient level), then the context is derived by (ctxIdxInc=((numDecodAbsLevelGt1 !=0) ? 0: Min(4, 1+numDecodAbsLevelEq1))); Otherwise, context is divided by (ctxIdxInc=5+Min(4−((ctxBlockCat==3) ? 1:0), numDecodAbsLevelGt1)). Context processing for the first bin in the absolute values of the coefficient minus one (i.e. Coeff level BinIdx 0 in FIG. 7 is different from the other bins in the coefficient level.
In one embodiment, the encoding Coeff Level BinIdx 0 occurs in a separate bin-plane as shown in the second row of FIG. 7 . The advantage in the context processing, because it can be carried out in parallel to the rest of the context processing i.e. the context processing for all the Coeff Level BinIdx 0 for all the coefficients level in a block, may be carried out in parallel to bin processing of Coeff Level BinIdx 0 before the decoding of the other bins in the coefficient level. This is referred to as Coeff Level BinIdx PCP.
In AVC, sign information is interleaved along with level information as shown in FIG. 5 . This leads to inefficiency in parallel context processing. FIG. 5 is an embodiment of a flow diagram depicting a method for context processing tree for level coding in AVC. In FIG. 5 , the context processing tree that needs to be pre-calculated at each bin to achieve 4× parallelism in context processing of level in AVC. The context processing that happens at every SIGN node is wasteful since SIGN is coded in bypass mode. Table 3 shows the distribution of level of coefficients obtained by measuring levels in bitstreams generated by, for example, a TMuC-0.1 using cfg files in cfp-fast directory.
TABLE 3 | |||
Probablity of | |||
| occurrence | ||
1 | 0.76 | ||
2 | 0.15 | ||
3 | 0.05 | ||
4 | 0.02 | ||
5 | 0.01 | ||
Level=1 occurs with the highest probability, so the most probable path in the context processing tree of FIG. 6 is L0(0)→SIGN0→L1(0)→SIGN1. For this particular path, the context processing efficiency is 50%, meaning half the context processing is wasteful. On the average, for the context processing tree of FIG. 6 . and assuming the level distribution of Table 3, the context processing efficiency is 60%. FIG. 6 is an embodiment of a flow diagram depicting context processing tree for levels when SIGN is coded in separate bin-plane. In FIG. 6 , the context processing tree for levels when sign is coded in separate bin-plane. As can be seen in FIG. 6 , context processing efficiency is 100%. This is also illustrated in FIG. 7 where all sign bins (i.e. Coeff Sign Bins) are coded on separate bin plans; this is referred to as Coeff Sign PCP.
In some embodiment, the first two bins in the coefficient level are context coded. The rest of the bins, such as, coefficient sign bins and Golomb-Rice+Exp-Golomb (GR-EG) binarized bins, are bypass coded. As an extension of “Coeff Level BinIdx 0 PCP”, the second bin in the absolute value of the coefficient minus 1 (i.e. Coeff Level BinIdx 1) is also coded in a separate bin-plane. The Coeff Sign Level can be interleaved or be on a separate bin-plane with GR-EG bins.
Since bypass coding is simpler than context coding, bypass bins can be coded faster than context coded bins. In particular, many bypass bins can be coded in a cycle which can increase the throughput of the CABAC. With Coeff Level BinIdx 1 PCP all bypass coded bins for coefficients in a given TU are grouped together which increases throughput impact of parallel bypass bins processing.
Variants of this approach include separating GR-EG +sign bins from the Coeff Level BinIdx 0 and Coeff Level BinIdx 1, but keeping the GR-EG +sign bins interleaved and keeping the Coeff Level BinIdx 0 and Coeff Level Bin Idx 1 bins interleaved as shown in proposal # 2 in FIG. 11 . This eliminates the additional of loops required to separate the Coeff Level BinIdx 0 from Coeff Level BinIdx 1, and Coefficient Sign from RG-EG bins. Alternatively, proposal # 3 in FIG. 7 keeps GR-EG and Coefficient sign bins interleaved, to reduce the loops, and keeps Coeff Level BinIdx 0 and Coeff Level BinIdx 1 in separate partitions. This is due to the fact that the context selection for Coeff Level BinIdx 0 and Coeff Level BinIdx 1 are more complex and keeping the two separate helps to improve parallel processing as described in section Coeff Level BinIdx 0 PCP. This approach can also be applied to other type of syntax elements such as motion vector difference. The bypass bins of the motion vectors difference can be coded together on a separate bin-plane than the context coded bins.
While the foregoing is directed to embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (16)
1. A method of encoding video information comprising:
receiving a series of syntax elements comprising an array of transform coefficients related to a block of an image;
determining a context for at least a first bin and a second bin of each significant transform coefficient in the array;
context encoding the at least a first bin and a second bin of each significant transform coefficient in the array;
bypass encoding a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array;
bypass encoding a sign indicator for each significant transform coefficient in the array;
constructing a bit stream comprising:
the context encoded first bins of each significant transform coefficient in the block;
the context encoded second bins of each significant transform coefficient in the block immediately following the context encoded first bins of each significant transform coefficient in the block;
the bypass encoded sign indicators for each significant transform coefficient in the array grouped together in the bit stream and after the first bin and second bin; and
the bypass encoded Golomb-Rice representations of the at least a most significant portion of at least one transform coefficient in the array, all Golomb-Rice representations for the array grouped together in the bit stream and after the first bin and second bin; and
transmitting the bit stream.
2. The method of claim 1 , comprising:
encoding header information into the bit steam before the context encoded first bin in the bit steam.
3. The method of claim 1 , comprising:
encoding a significance map into a header portion of the bit stream located before the context encoded first bin in the bit steam.
4. The method of claim 1 , comprising:
encoding motion vector information into a header portion of the bit stream located before the context encoded first bin in the bit steam.
5. The method of claim 1 , comprising:
encoding prediction information into a header portion of the bit stream located before the context encoded first bin in the bit steam.
6. The method of claim 1 , bypass encoding a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array only performed for transform coefficients that are not completely magnitude encoded by context encoding.
7. The method of claim 1 , the bypass encoded sign indicators for each significant transform coefficient in the array grouped together in the bit stream after all of the bypass encoded Golomb-Rice representations of transform coefficients in the array.
8. An encoder comprising:
a binarizer operable to binarize syntax elements;
a context modeler operable to determine a context for a portion of the binarized syntax elements, including at least a first bin and a second bin of each significant transform coefficient in an array of transform coefficients related to a block of an image;
a binary arithmetic coder comprising:
a context coding engine operable to context encode the at least a first bin and a second bin of each significant transform coefficient in the array; and
a bypass coding engine operable to:
bypass encode a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array; and
bypass encode a sign indicator for each significant transform coefficient in the array;
the binary arithmetic coder operable to group coded elements to form a bit stream having:
a group of context encoded first bins of each significant transform coefficient in the block;
a group of context encoded second bins of each significant transform coefficient in the block immediately following the group of context encoded first bins of each significant transform coefficient in the block;
a group of bypass encoded sign indicators for each significant transform coefficient in the array in the bit stream and after the group of context encoded first bins and the group of context encoded second bins; and
a group of bypass encoded Golomb-Rice representations of the at least a most significant portion of at least one transform coefficient in the array in the bit stream and after the group of context encoded first bins and the group of context encoded second bins.
9. The encoded of claim 8 , the binary arithmetic coder further operable to encode header information into the bit steam before the context encoded first bin in the bit steam.
10. The encoded of claim 8 the binary arithmetic coder further operable to encode a significance map into a header portion of the bit stream located before the context encoded first bin in the bit steam.
11. The encoded of claim 8 the binary arithmetic coder further operable to encode a significance map into a header portion of the bit stream only when at least one transform coefficient in the array is non-zero.
12. The encoded of claim 8 the binary arithmetic coder further operable to encode motion vector information into a header portion of the bit stream located before the context encoded first bin in the bit steam.
13. The encoded of claim 8 the binary arithmetic coder further operable to encode prediction information into a header portion of the bit stream located before the context encoded first bin in the bit steam.
14. The encoded of claim 8 the bypass coding engine operable to bypass encode a Golomb-Rice representation of an additional portion of at least one transform coefficient in the array only for transform coefficients that are not completely magnitude encoded by context encoding.
15. The encoded of claim 8 the binary arithmetic coder operable to group encoded sign indicators for each significant transform coefficient in the array together in the bit stream after all of the bypass encoded Golomb-Rice representations of transform coefficients in the array.
16. A method of encoding video information comprising:
receiving a series of syntax elements comprising an array of transform coefficients related to a block of an image;
determining a context for at least a first bin and a second bin of each significant transform coefficient in the array;
context encoding the at least a first bin and a second bin of each significant transform coefficient in the array;
bypass encoding a Golomb-Rice representation of an additional portion of any transform coefficients that are not completely magnitude encoded by the context encoding;
bypass encoding a sign indicator for each significant transform coefficient in the array;
constructing a bit stream comprising:
a header portion comprising a significance map, motion vector information, and prediction information
the context encoded first bins of each significant transform coefficient in the block following the header portion;
the context encoded second bins of each significant transform coefficient in the block immediately following the context encoded first bins of each significant transform coefficient in the block;
the bypass encoded sign indicators for each significant transform coefficient in the array grouped together in the bit stream and after the first bin and second bin; and
the bypass encoded Golomb-Rice representations of the at least a most significant portion of at least one transform coefficient in the array, all Golomb-Rice representations for the array grouped together in the bit stream and after the first bin and second bin; and
transmitting the bit stream.
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US13/184,226 US9591320B2 (en) | 2010-07-15 | 2011-07-15 | Context and bypass encoding video |
US15/295,689 US10362322B2 (en) | 2010-07-15 | 2016-10-17 | Context and bypass encoding video |
US16/518,407 US11490103B2 (en) | 2010-07-15 | 2019-07-22 | Context and bypass encoding video |
US16/891,353 US10939131B2 (en) | 2010-07-15 | 2020-06-03 | Context and bypass encoding video |
US17/936,003 US11750826B2 (en) | 2010-07-15 | 2022-09-28 | Context and bypass encoding video |
US18/241,316 US20230412826A1 (en) | 2010-07-15 | 2023-09-01 | Context and bypass encoding video |
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US36459310P | 2010-07-15 | 2010-07-15 | |
US201161499852P | 2011-06-22 | 2011-06-22 | |
US13/184,226 US9591320B2 (en) | 2010-07-15 | 2011-07-15 | Context and bypass encoding video |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US15/295,689 Continuation US10362322B2 (en) | 2010-07-15 | 2016-10-17 | Context and bypass encoding video |
Publications (2)
Publication Number | Publication Date |
---|---|
US20120014454A1 US20120014454A1 (en) | 2012-01-19 |
US9591320B2 true US9591320B2 (en) | 2017-03-07 |
Family
ID=45466981
Family Applications (6)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US13/184,226 Active 2034-06-28 US9591320B2 (en) | 2010-07-15 | 2011-07-15 | Context and bypass encoding video |
US15/295,689 Active US10362322B2 (en) | 2010-07-15 | 2016-10-17 | Context and bypass encoding video |
US16/518,407 Active US11490103B2 (en) | 2010-07-15 | 2019-07-22 | Context and bypass encoding video |
US16/891,353 Active US10939131B2 (en) | 2010-07-15 | 2020-06-03 | Context and bypass encoding video |
US17/936,003 Active US11750826B2 (en) | 2010-07-15 | 2022-09-28 | Context and bypass encoding video |
US18/241,316 Pending US20230412826A1 (en) | 2010-07-15 | 2023-09-01 | Context and bypass encoding video |
Family Applications After (5)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US15/295,689 Active US10362322B2 (en) | 2010-07-15 | 2016-10-17 | Context and bypass encoding video |
US16/518,407 Active US11490103B2 (en) | 2010-07-15 | 2019-07-22 | Context and bypass encoding video |
US16/891,353 Active US10939131B2 (en) | 2010-07-15 | 2020-06-03 | Context and bypass encoding video |
US17/936,003 Active US11750826B2 (en) | 2010-07-15 | 2022-09-28 | Context and bypass encoding video |
US18/241,316 Pending US20230412826A1 (en) | 2010-07-15 | 2023-09-01 | Context and bypass encoding video |
Country Status (1)
Country | Link |
---|---|
US (6) | US9591320B2 (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200145681A1 (en) * | 2018-11-02 | 2020-05-07 | Fungible, Inc. | Memory layout for jpeg accelerator |
US10827192B2 (en) | 2018-11-02 | 2020-11-03 | Fungible, Inc. | Work allocation for JPEG accelerator |
US10827191B2 (en) | 2018-11-02 | 2020-11-03 | Fungible, Inc. | Parallel coding of syntax elements for JPEG accelerator |
US10931958B2 (en) | 2018-11-02 | 2021-02-23 | Fungible, Inc. | JPEG accelerator using last-non-zero (LNZ) syntax element |
US11917209B2 (en) * | 2018-09-24 | 2024-02-27 | Tencent America LLC | Context model reduction for transform coefficients entropy coding |
Families Citing this family (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120230418A1 (en) | 2011-03-08 | 2012-09-13 | Qualcomm Incorporated | Coding of transform coefficients for video coding |
US9106913B2 (en) * | 2011-03-08 | 2015-08-11 | Qualcomm Incorporated | Coding of transform coefficients for video coding |
US8855207B2 (en) | 2011-06-24 | 2014-10-07 | Panasonic Intellectual Property Corporation Of America | Image coding method for encoding a difference motion vector into a bitstream |
PL2728864T3 (en) * | 2011-06-28 | 2018-05-30 | Nec Corporation | Method for coding video quantization parameter and method for decoding video quantization parameter |
MX2013001773A (en) * | 2011-07-18 | 2013-04-03 | Panasonic Corp | Image encoding method, image decoding method, image encoding apparatus, image decoding apparatus, and image encoding/decoding apparatus. |
US9154792B2 (en) | 2011-11-08 | 2015-10-06 | Qualcomm Incorporated | Progressive coding of position of last significant coefficient |
US20130114667A1 (en) * | 2011-11-08 | 2013-05-09 | Sony Corporation | Binarisation of last position for higher throughput |
US9357185B2 (en) | 2011-11-08 | 2016-05-31 | Qualcomm Incorporated | Context optimization for last significant coefficient position coding |
WO2013082291A2 (en) * | 2011-11-29 | 2013-06-06 | Huawei Technologies Co., Ltd. | Unified partitioning structures and signaling methods for high efficiency video coding |
US9503717B2 (en) * | 2012-01-09 | 2016-11-22 | Texas Instruments Incorporated | Context adaptive binary arithmetic coding (CABAC) with scalable throughput and coding efficiency |
WO2013106987A1 (en) * | 2012-01-16 | 2013-07-25 | Mediatek Singapore Pte. Ltd. | Methods and apparatuses of bypass coding and reducing contexts for some syntax elements |
AU2012200345B2 (en) * | 2012-01-20 | 2014-05-01 | Canon Kabushiki Kaisha | Method, apparatus and system for encoding and decoding the significance map residual coefficients of a transform unit |
US9161035B2 (en) | 2012-01-20 | 2015-10-13 | Sony Corporation | Flexible band offset mode in sample adaptive offset in HEVC |
KR102097668B1 (en) | 2012-01-20 | 2020-04-06 | 지이 비디오 컴프레션, 엘엘씨 | Transform coefficient coding |
KR101669524B1 (en) | 2012-02-01 | 2016-11-09 | 노키아 테크놀로지스 오와이 | Method and apparatus for video coding |
US9237344B2 (en) * | 2012-03-22 | 2016-01-12 | Qualcomm Incorporated | Deriving context for last position coding for video coding |
CN107197283B (en) * | 2012-03-26 | 2020-06-12 | Jvc 建伍株式会社 | Image encoding device, image encoding method, and transmission device |
CN108134942B (en) * | 2012-04-11 | 2020-09-15 | 杜比国际公司 | Method for encoding and decoding a bitstream associated with transform coefficients |
US9538172B2 (en) | 2012-04-11 | 2017-01-03 | Qualcomm Incorporated | Grouping bypass coded syntax elements in video coding |
TWI527466B (en) | 2012-04-13 | 2016-03-21 | Ge影像壓縮有限公司 | Low delay picture coding |
WO2013175736A1 (en) * | 2012-05-25 | 2013-11-28 | パナソニック株式会社 | Video encoding method, video encoding device, video decoding method, video decoding device, and video encoding/decoding device |
JP6179813B2 (en) | 2012-05-25 | 2017-08-16 | サン パテント トラスト | Image encoding method, image encoding device, image decoding method, image decoding device, and image encoding / decoding device |
KR102132917B1 (en) | 2012-05-25 | 2020-07-10 | 선 페이턴트 트러스트 | Video image coding method, video image decoding method, video image coding device, video image decoding device, and video image coding-decoding device |
CN103650498B (en) | 2012-06-04 | 2018-01-05 | 太阳专利托管公司 | Dynamic image decoding method and moving image decoding apparatus |
TWI737990B (en) | 2012-06-29 | 2021-09-01 | 美商Ge影像壓縮有限公司 | Video data stream concept |
US9936200B2 (en) * | 2013-04-12 | 2018-04-03 | Qualcomm Incorporated | Rice parameter update for coefficient level coding in video coding process |
US10021419B2 (en) | 2013-07-12 | 2018-07-10 | Qualcomm Incorported | Rice parameter initialization for coefficient level coding in video coding process |
US10158874B2 (en) * | 2015-09-30 | 2018-12-18 | Apple Inc. | Parallel bypass and regular bin coding |
US10554977B2 (en) * | 2017-02-10 | 2020-02-04 | Intel Corporation | Method and system of high throughput arithmetic entropy coding for video coding |
US10506258B2 (en) * | 2017-07-13 | 2019-12-10 | Google Llc | Coding video syntax elements using a context tree |
CN107580224B (en) * | 2017-08-08 | 2019-11-22 | 西安理工大学 | A kind of adaptive scanning method towards HEVC entropy coding |
US10992937B2 (en) * | 2018-06-18 | 2021-04-27 | Qualcomm Incorporated | Coefficient coding with grouped bypass bins |
US11113846B2 (en) * | 2018-08-31 | 2021-09-07 | Hulu, LLC | Coefficient context modeling in video coding |
US20230059794A1 (en) * | 2021-08-23 | 2023-02-23 | Mediatek Inc. | Context-based adaptive binary arithmetic coding decoder capable of decoding multiple bins in one cycle and associated decoding method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090074057A1 (en) * | 2004-01-30 | 2009-03-19 | Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. | Video frame encoding and decoding |
US20090096643A1 (en) * | 2007-10-16 | 2009-04-16 | Yuan-Teng Chang | System and Method for Context-Based Adaptive Binary Arithematic Encoding and Decoding |
US20110243226A1 (en) * | 2010-04-05 | 2011-10-06 | Samsung Electronics Co., Ltd. | Low complexity entropy-encoding/decoding method and apparatus |
US20120027089A1 (en) * | 2010-07-28 | 2012-02-02 | Qualcomm Incorporated | Coding motion vectors in video coding |
US20120230417A1 (en) * | 2011-03-08 | 2012-09-13 | Qualcomm Incorporated | Coding of transform coefficients for video coding |
US8660177B2 (en) * | 2010-03-24 | 2014-02-25 | Sony Computer Entertainment Inc. | Parallel entropy coding |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5717394A (en) * | 1993-02-10 | 1998-02-10 | Ricoh Company Ltd. | Method and apparatus for encoding and decoding data |
US7710296B2 (en) * | 2007-09-19 | 2010-05-04 | Texas Instruments Incorporated | N-bin arithmetic coding for context adaptive binary arithmetic coding |
US8294603B2 (en) * | 2009-06-30 | 2012-10-23 | Massachusetts Institute Of Technology | System and method for providing high throughput entropy coding using syntax element partitioning |
-
2011
- 2011-07-15 US US13/184,226 patent/US9591320B2/en active Active
-
2016
- 2016-10-17 US US15/295,689 patent/US10362322B2/en active Active
-
2019
- 2019-07-22 US US16/518,407 patent/US11490103B2/en active Active
-
2020
- 2020-06-03 US US16/891,353 patent/US10939131B2/en active Active
-
2022
- 2022-09-28 US US17/936,003 patent/US11750826B2/en active Active
-
2023
- 2023-09-01 US US18/241,316 patent/US20230412826A1/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090074057A1 (en) * | 2004-01-30 | 2009-03-19 | Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V. | Video frame encoding and decoding |
US20090096643A1 (en) * | 2007-10-16 | 2009-04-16 | Yuan-Teng Chang | System and Method for Context-Based Adaptive Binary Arithematic Encoding and Decoding |
US8660177B2 (en) * | 2010-03-24 | 2014-02-25 | Sony Computer Entertainment Inc. | Parallel entropy coding |
US20110243226A1 (en) * | 2010-04-05 | 2011-10-06 | Samsung Electronics Co., Ltd. | Low complexity entropy-encoding/decoding method and apparatus |
US20120027089A1 (en) * | 2010-07-28 | 2012-02-02 | Qualcomm Incorporated | Coding motion vectors in video coding |
US20120230417A1 (en) * | 2011-03-08 | 2012-09-13 | Qualcomm Incorporated | Coding of transform coefficients for video coding |
Non-Patent Citations (9)
Title |
---|
Andrew Segall, et al., Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11 "A Highly Efficient and Highly Parallel System for Video Coding" Dresden, DE, Apr. 15-23, 2010. |
D. He, et al., Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11 "Video Coding Technology Proposal by RIM" Dresden, DE, Apr. 15-23, 2010. |
Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11 "Test Model under Consideration" Dresden, DE, Apr. 15-23, 2010. |
M. Budagavi, V. Sze, et al., Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11 "Description of Video Coding Technology Proposal by Texas Instruments, Inc." Dresden, DE, Apr. 15-23, 2010. |
Sze et al., "A high throughput CABAC algorithm using syntax element partitioning", Nov. 10, 2009, IEEE, 2009 16th IEEE International Conference on Image Processing. * |
T. Nguyen, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WPE and ISO/IEC JTC1/SC29/WG11 "CE 11: Coding of Transform Coefficient Levels with Golomb-Rice Codes" Geneva, CH, Mar. 16-23, 2011. |
T. Weigand, et al., Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11 "Description of Video Coding Technology Proposal by Fraunhofer HHI" Dresden, DE, Apr. 15-23, 2010. |
Thomas Wiegand et al., "WD2: Working Draft 2 of High-Efficiency Video Coding", JCTVC-D503, Jan. 20-28, 2011, 1-555, Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, Daegu, KR. |
V. Sze, M. Budagavi, "Parallel CABAC," ITU-T SG 16/Q.6 Doc. COM16-C334, Geneva, Switzerland, Apr. 2008. |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11917209B2 (en) * | 2018-09-24 | 2024-02-27 | Tencent America LLC | Context model reduction for transform coefficients entropy coding |
US20200145681A1 (en) * | 2018-11-02 | 2020-05-07 | Fungible, Inc. | Memory layout for jpeg accelerator |
US10827192B2 (en) | 2018-11-02 | 2020-11-03 | Fungible, Inc. | Work allocation for JPEG accelerator |
US10827191B2 (en) | 2018-11-02 | 2020-11-03 | Fungible, Inc. | Parallel coding of syntax elements for JPEG accelerator |
US10848775B2 (en) * | 2018-11-02 | 2020-11-24 | Fungible, Inc. | Memory layout for JPEG accelerator |
US10931958B2 (en) | 2018-11-02 | 2021-02-23 | Fungible, Inc. | JPEG accelerator using last-non-zero (LNZ) syntax element |
Also Published As
Publication number | Publication date |
---|---|
US11750826B2 (en) | 2023-09-05 |
US20200304795A1 (en) | 2020-09-24 |
US20230028157A1 (en) | 2023-01-26 |
US20230412826A1 (en) | 2023-12-21 |
US20120014454A1 (en) | 2012-01-19 |
US10362322B2 (en) | 2019-07-23 |
US10939131B2 (en) | 2021-03-02 |
US20190342566A1 (en) | 2019-11-07 |
US11490103B2 (en) | 2022-11-01 |
US20170034515A1 (en) | 2017-02-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11750826B2 (en) | Context and bypass encoding video | |
JP7100836B2 (en) | Context initialization in entropy coding | |
US9838034B2 (en) | Method and apparatus for reordered binarization of syntax elements in CABAC | |
US7564384B2 (en) | Binarizing method and device thereof | |
US20100097248A1 (en) | Method and apparatus for video processing in context-adaptive binary arithmetic coding | |
KR102393305B1 (en) | Entropy encoding and decoding scheme | |
US20080246637A1 (en) | Cabac Decoding Method | |
US20070139236A1 (en) | Variable Length Coding Algorithm for Multiple Coding Modes | |
EP2381686B1 (en) | Apparatus for parallel entropy encoding and decoding | |
Budagavi et al. | Parallel Context Processing techniques for high coding efficiency entropy coding in HEVC. Status: Input Document to JCT-VC Purpose: Proposal | |
Tian et al. | Implementation strategies for statistical codec designs in h. 264/avc standard |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: TEXAS INSTRUMENTS INCORPORATED, TEXAS Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BUDAGAVI, MADHUKAR;DEMIRCIN, MEHMET UMUT;SZE, VIVIENNE;SIGNING DATES FROM 20110712 TO 20110818;REEL/FRAME:027079/0204 |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 4 |